Evaluation of Possible Reproductively Mediated Character Displacement in the Crayfishes, Orconectes Rusticus and 0

Evaluation of Possible Reproductively Mediated Character Displacement in the Crayfishes, Orconectes rusticus and 0. sanbornii1 MARK J. BUTLER IV, Department of Biological Sciences, The Florida State University, Tallahassee, FL 323O6-3O5O ABSTRACT. Orconectes rusticus is replacing several species of crayfishes in north-central and northeastern North America, including O. sanbornii in Ohio. Recent evidence suggests that the species replacements may be driven by asymmetrical reproductive success favoring 0. rusticus. Nonetheless, some sympatric associations appear locally persistent. Because crayfish demonstrate size-assortative mating and there is a disparity in the sizes of the species, further divergence in the sizes of the species in sympatry could enhance reproductive isolation, ultimately providing a mechanism for character displacement. To test this hypothesis the size differentials between crayfish collected from allopatric and sympatric populations in east-central Ohio were compared with expected differences. The possibility of clinal variation in size was addressed by comparing crayfish sizes along a continuous allopatric-sympatric-allopatric species gradient within one stream. The initial character displacement hypothesis was not substantiated by comparisons of allopatric and sympatric populations within or among streams for male or female 0. rusticus or male 0. sanbornii. However, female 0. sanbornii size distributions were consistent with unilateral character displacement. OHIO J. SCI. 88 (3): 87-91, 1988 INTRODUCTION that these deviations were consistent with character dis- The classic definition of character displacement placement. Orconectes rusticus was introduced into the (Brown and Wilson 1956) involves two closely related Licking River drainage (Knox and Licking counties, species with overlapping geographical distributions that Ohio) sometime after Turner's (1926) initial crayfish sur- are less similar in sympatry than allopatry. The original vey, but before 1967 (R. Jezerinac, unpubl. data). Thus explanation for character displacement centered on two the species have been interacting for at least 10 genera- mechanisms: interspecific competition or interspecific tions, a short but demonstrably sufficient time for the mating coupled with inviable hybrid offspring. Either evolution of a variety of traits in many animal groups process could favor the divergence of various key traits. (Doyle and Hunte 1981, Reznick 1982, Rice 1985, Grant (1972) later described character displacement in Endler 1986). Although 0. sanbornii had been extirpated more detail. Although he acknowledged the potential for from some stream sites by 0. rusticus since the last census reproductively based species divergence, he concentrated of the area (east-central Ohio; R. Jezerinac, unpubl. on competitive hypotheses. This perspective has per- data), many sympatric populations persisted. Similar situ- vaded the study of character displacement since then ations have been reported in southern Ohio (Flynn and (Slatkin 1980, Strong et al. 1984, Taper and Case 1985). Hobbs 1984), northern Wisconsin (Capelli 1982, Capelli Much of the recent criticism that research on character and Magnuson 1983, Lodge et. al. 1986), and Ontario displacement has received centers on the paucity of aut- (Berrill 1978, Tierney and Dunham 1984). ecological data available to substantiate claims of com- Rapid evolution of divergent traits requires strong petition and on the inability of many studies to reject the selection. Phenotypic correlations between species, par- alternative hypotheses of clinal variation or allopatric ticularly from only a few sites, cannot be accepted as character release (Grant 1972, Hespenheide 1973, Strong evidence for character displacement (Endler 1985). Thus et al. 1979, 1984, Grant and Abbott 1980, Strong and the existence of a sufficiently powerful selective agent Simberloff 1981). These objections are untenable in the must also be demonstrated. Several independent studies system described in this paper because 1) interspecific on crayfish mating and reproduction suggest that such a interactions driving species replacement and possibly mechanism may exist (Capelli and Capelli 1980, Capelli character displacement are documented and apparently 1982, Tierney and Dunham 1982, 1984, Berrill 1985, do not involve competition/WJ?; 2) allopatric-sympatric Butler and Stein 1985, Butler 1985, Lodge et al. 1985, population gradients are localized and are thus not sub- 1986). Orconectes rusticus males rarely engage in hetero- ject to geographical clines; and 3) sympatric associations specific copulations, whereas male 0. sanbornii and 0. are more recent than the allopatric condition, negating propinquus (a species closely related to 0. sanbornii and the possibility of character release. native to Wisconsin and Canada) frequently mate hetero- In a previous study I investigated the replacement of specifically with small 0. rusticus females (Tierney and a native Ohio crayfish, Orconectes sanbornii, by an intro- Dunham 1984, Butler and Stein 1985, Berrill 1985, but duced crayfish, Orconectes rusticus, and evaluated several see Capelli and Capelli 1980, Capelli 1982). Females that mechanisms that might drive species replacements (But- mate heterospecifically experience reductions in re- ler and Stein 1985). During that study we noted that the productive success of 50 to 90% (Berrill 1985, Butler and size ratios (i.e., carapace lengths and chelae lengths) of Stein 1985). Although the selection of small 0. rusticus the species were more disparate in sympatry (1.29 • 1; 0. females by 0. sanbornii (or 0. propinquus) males lowers rusticus vs. 0. sanbornii) than in allopatry (1.13: 1), and recruitment for all species, the relative recruitment of 0. rusticus is potentially greater. This probably occurs because only large males (i.e., 0. rusticus) can copulate with 'Manuscript received 4 March 1987 and in revised form 31 August large 0. rusticus females (Tierney and Dunham 1984, 1987 (#87-13). M.J. BUTLER IV Vol. 88 Berrill and Arsenault 1982, 1984), and large females are 34 generally more fecund than small females (Fielder 1972, MALES Lorman 1980, Hazlett 1983, Salmon 1983). A mating 32- system of this type should precipitate a decline in the population size of the native species (0. sanbornii or 0. propinquus) unless the populations are subject to substan- tial immigration or the conflicting species become more ) reproductively isolated. Sympatric divergence in the relative size of the two (MM species is one possible isolating mechanism. Most hetero- H specific matings should occur between small 0. rusticus females and 0. sanbornii males because orconectid crayfish mate with individuals of similar size (Berrill and Arsenault 1982, 1984), 0. rusticus is generally larger than LENGT 0. sanbornii (Turner 1926, Butler and Stein 1985), and E FEMALES 0. rusticus males demonstrate greater species-specific mate selectivity (Tierney and Dunham 1984, Butler and Stein 1985). Individuals that mate heterospecifically will experience lowered reproductive success relative to other individuals in sympatric populations because their fe- CARAPAC cundity is reduced and hybrids are apparently nonviable (Capelli and Capelli 1980, Smith 1981, Berrill 1985, Butler and Stein 1985). Thus, relatively larger 0. rusticus 24- and smaller 0. sanbornii might be selectively favored in sympatry, leading to character displacement in body size 22- and other highly correlated characters like chelae length 20 (Stein 1975, Stein et al. 1977). I tested this hypothesis 12 3456789 10 11 by examining the allopatric and sympatric size distributions of two crayfish species within and among several STREAM SITE Ohio streams. FIGURE 1. Mean carapace lengths for male and female Orconectes METHODS rusticus (solid squares) and Orconectes sanbornii (open squares) at 11 sites along a 3-km transect of the North Fork Licking River, Licking and I visited a total of 32 stream sites in northern Licking and Knox Knox counties, Ohio. Note that the transect spanned two allopatric counties in 1984 and located only two 0. rustkus-0. sanbornii sym- zones and a zone of sympatry in between. Twenty-five to 35 crayfish patric populations and four allopatric 0. rusticus populations. The were collected at each site. remaining sites yielded only 0. sanbornii populations. Therefore, crayfish were collected from the two sympatric, the four allopatric 0. rusticus, and five nearby allopatric 0. sanbornii locales in five greater than 20 mm CL (i.e., those capable of reproduction; Lorman east-central Ohio streams in July, 1984. All five streams sampled are 1980, Fielder 1972, Berrill and Arsenault 1984) were used in the low gradient, third-order streams that drain agricultural watersheds. analyses. Orconectes rusticus was collected at four allopatric sites in the Otter Size-frequency data were first analyzed in one-factor, fixed-effects Fork, Muddy Fork, and North Fork of the Licking River and in analyses of variance (ANOVA; factor = collection site) to determine Raccoon Creek. Orconectes sanbornii was collected at five allopatric sites if sizes differed among sites. Bonferroni a priori multiple comparison in the North Fork and South Forks of the Licking River; both spe- tests (Kirk 1982) were then used to test whether crayfish size differed cies were taken at two sympatric locations in the North Fork of the between allopatric and sympatric sites. I also used Wilcoxon Rank- Licking River. Sum tests (Hollander and Wolfe

Load more